MozReview Statistics July 2015
July 16, 2015 at 02:00 PM | categories: MozReview, MozillaAs of today, ~15.6% of commits landing in Firefox in July have gone through MozReview or have been produced on machines that have used MozReview. This is still a small percentage of overall commits. But, signs are that the percentage is going up. Last month, about half as many commits exhibited the same signature. It's only July 16 and we've already passed the total from June.
What I find interesting is the differences between commits that have gone through MozReview versus the rest. When you look at the diff statistics (a quick proxy of change size), we find that MozReview commits tend to be smaller. The median adds as reported by diff stat (basically lines that were changed) is 12 for MozReview versus 17 elsewhere. The average is 58 for MozReview versus 100 elsewhere. For number of files modified, MozReview averages 2.59 versus elsewhere's 2.71. (These numbers exclude some specific large commits that appeared to be bulk imports of external projects and drove up the non-MozReview figures.)
It's entirely possible the root cause behind the discrepancy is a side-effect of the population of MozReview users: perhaps MozReview users just write smaller commits. However, I'd like to think it's because MozReview makes it easier to manage multiple commits and people are taking advantage of that (this is an explicit design goal of MozReview). Whatever the root cause, I'm glad diffs are smaller. As I've written about before, smaller commits are easier to review and land, thus enabling projects to move faster.
I have a quarterly goal to remove the requirement for a Mozilla LDAP account to push to MozReview. That will allow first time contributors to use MozReview. This will be a huge win, as we can do much more magic in the MozReview world than we can from vanilla Bugzilla (automatic bug filing, automatic reviewer assignment, etc). Unofficially, I'd like to have more than 50% of Firefox commits go through MozReview by the end of the year.
Cloning From S3
July 08, 2015 at 11:40 AM | categories: Mercurial, MozillaA month ago, I blogged about faster cloning from hg.mozilla.org using bundle files. But deploying the feature on the servers was only the tip of the iceberg.
At the end of last week, Firefox release automation rolled out the bundleclone extension to their Linux and OS X machines. Essentially, clones are bootstrapped from Amazon S3 automatically once this extension is installed.
In just a few days of deployment, we've already seen a drastic shift in traffic. On UTC day 2015-07-07, S3 served 1,563,014,396,236 bytes of repository data! The hg.mozilla.org servers themselves served a total of 1,976,057,201,583 bytes.
But we're not done. The bundleclone extension isn't yet deployed on Windows. Nor is it always used on TaskCluster. In addition, there are still some high-use repositories that don't have bundles being generated. Yesterday, I crunched the data and enabled bundles on more repositories. It's too early to have conclusive data, but this should move an additional several hundred gigabytes of traffic to S3.
At Whistler, I said that reducing traffic on hg.mozilla.org by 90% is within the realm of possibility. Between a partial rollout of S3 clone offload that is already serving 44% of traffic and a feature I'm working on in core Mercurial to enable auto sharing of repository data, I'd say we're well on track.
Publish When Pushing to MozReview
July 07, 2015 at 02:55 PM | categories: MozReview, MozillaA lot of people contributed some really great feedback about MozReview at Whistler. One of the most frequent requests was for the ability to publish submitted review requests without having to open a browser. I'm pleased to report that as of yesterday, this feature is implemented! If reviewers have been assigned to all your review requests, Mercurial will now prompt you to publish the review requests during hg push. It should just work.
As part of this change, we also introduced more advanced feature negotiation into the handshake between client and server. This means we now have a mechanism for detecting out-of-date client installations. This will enable us to more aggressively drop backwards compatibility (making server-side development easier) while simultaneously ensuring that more people are running modern and hopefully better versions of the client code. This should translate to moving faster and a better experience for everyone.
Changeset Metadata on hg.mozilla.org
June 04, 2015 at 01:55 PM | categories: Mercurial, MozillaJust a few minutes ago, I deployed some updates to hg.mozilla.org to display more metadata on changeset pages. See 4b69a62d1905, dc4023d54436, and b617a57d6bf1 for examples of what's shown.
We currently display:
- More detailed pushlog info. (Before you had to load another page to see things like the date of the push.)
- The list of reviewers, each being a link that searches for other changesets they've reviewed.
- A concise list of bugs referenced in the commit message.
- Links to changesets that were backed out by this changeset.
- On changesets that were backed out, we may also display a message that the changeset was backed out.
- For Firefox repos, we also display the application milestone. This is the Gecko/app version recorded in the config/milestone.txt file in that changeset. The value can be used to quickly answer the question What versions of Firefox have this changeset.
If you notice any issues or have requests for new features, please file a bug.
This work is built on top of a feature I added to Mercurial 3.4 to make it easier to inject extra data into Mercurial's web templates. We just deployed Mercurial 3.4.1 to hg.mozilla.org yesterday. It's worth noting that this deployment happened in the middle of the day with no user-perceived downtime. This is a far cry from where we were a year ago, when any server change required a maintenance window. We've invested a lot of work into a test suite for this service so we can continuously deploy without fear of breaking things. Moving fast feels so good.
Important Changes to MozReview
May 29, 2015 at 04:20 PM | categories: MozReview, Mozilla, code reviewThis was a busy week for MozReview! There are a number of changes people need to be aware of.
Support for Specifying Reviewers via Commit Messages
MozReview will now parse r?gps syntax out of commit messages to set reviewers for pushed commits.
Read the docs for more information, including why we are preferring r? to r=.
Since it landed, a number of workflow concerns have been reported. See the bug tree for bug 1142251 before filing a bug to help avoid duplicates.
Thank Dan Minor for the feature!
Review Attachment/Flag Per Commit
Since the beginning of MozReview, there was one Bugzilla attachment / review flag per commit series. This has changed to one attachment / review flag per commit.
Before, you needed to grant Ship It on the parent/root review request in order to r+ the MozReview review request. Now, you grant Ship It on individual commits and these turn into individual r+ on Bugzilla. To reinforce that reviewing the parent/root review request doesn't do anything meaningful any more, the Ship It button and checkbox have been removed from the parent/root review request.
The new model more closely maps to how code review in Bugzilla has worked at Mozilla for ages. And, it is a superior workflow for future workflows we're trying to enable.
We tried to run a one-time migration script to convert existing parent/root attachments/review flags to per-commit attachments/flags. However, there were issues. We will attempt again sometime next week. In the interim, in-flight review requests may enter an inconsistent state if they are updated. If a new push is performed, the old parent/root attachment/review flag may linger and per-commit attachments/flags will be created. This could be confusing. The workaround is to manually clear the r? flag from the parent/root attachment or wait for the migration script to run in a few days.
Mark Côté put in a lot of hard work to make this change happen.
r? Flags Cleared After Review
Before, submitting a review without granting Ship It wouldn't do anything to the r? flag: the r? flag would linger.
Now, submitting review without granting Ship It will clear the r? flag. We think the new default is better for the majority of users. However, we recognize it isn't always wanted. There is a bug open to provide a checkbox to keep the r? flag present.
Metadata Added to Changesets
If you update to the tip of the version-control-tools repository (you should do this every week or so to stay fresh - use mach mercurial-setup to do this automatically), metadata will automatically be added to commits when working with MozReview-enabled repositories.
Specifically, we insert a hidden, unique, random ID into every changeset. This ID can be used to map commits to each other. We don't use this ID yet. But we have plans.
A side-effect of this change is that you can no longer push to MozReview if you have uncommitted local changes. If this is annoying, use hg shelve and hg unshelve to create and undo temporary commits. If this is too annoying, complain by filing a bug and we can look into doing this automatically as part of pushing.
What's Next?
We're actively working on more workflow enhancements to make MozReview an even more compelling experience.
I'm building a web service to query file metadata from moz.build files. This will allow MozReview to automatically file bugs in proper components based on what files changed. Once code reviewer metadata is added to moz.build files, it will enable us to assign reviewers automatically as well. The end goal here is to lower the number of steps needed to turn changed code into a landing. This will be useful when we turn GitHub pull requests into MozReview review requests (random GitHub contributors shouldn't need to know who to flag for review, nor should they be required to file a bug if they write some code). Oh year, we're working on integrating GitHub pull requests.
Another area of focus is better commit tracking and partially landed series. I have preliminary patches for automatically adding review URL annotations to commit messages. This will enable people to easily go from commit (message) to MozReview, without having to jump through Bugzilla. This also enables better commit tracking. If you e.g. perform complicated history rewriting, the review URL annotation will enable the MozReview server to better map previously-submitted commits to existing review requests. Partially landed series will enable commits to land as soon as they are reviewed, without having to wait on the entire series. It's my strong belief that if a commit is granted review, it should land as soon as possible. This helps prevent bit rot of ready-to-land commits. Landings also make people feel better because you feel like you've accomplished something. Positive feedback loops are good.
Major work is also being done to overhaul the web UI. The commit series view (which is currently populated via XHR) will soon be generated on the server and served as part of the page. This should make pages load a bit faster. And, things should be prettier as well.
And, of course, work is being invested into Autoland. Support for submitting pushes to Try landed a few weeks ago. Having Autoland actually land reviewed commits is on the radar.
Exciting times are ahead. Please continue to provide feedback. If you see something, say something.
« Previous Page -- Next Page »